“The hook was weak” is the least useful note in short-form. It is true of almost every underperforming video and it points at nothing. These are the four failures worth separating, because each has a different fix.
1. There is no gap
The opening states the topic and stops. Today I want to talk about retention. Nothing has been promised, so nothing is pending. The viewer knows what the video is about and has no reason to stay for the part where it becomes about something.
The fix is not more energy. Delivering a flat opening loudly is still a flat opening. The fix is to move a consequence forward — state the thing the topic leads to, and let the topic arrive as explanation.
2. The gap is real, but nobody has it
This is the expert's failure. The opening assumes a question the audience has never asked, usually because the creator has spent months in a problem the viewer encountered yesterday for the first time. Why trimmed medians beat raw averages is a genuine gap for about four hundred people.
The fix is to find the version of the question your audience already has. Usually it is one level less sophisticated and phrased as a complaint rather than a question.
3. The gap opens and never closes
The hook promises something the video does not deliver, either because the payoff never arrives or because it arrives so diluted that nobody registers it as the answer. This is the failure that does long-term damage: it performs acceptably once and degrades every subsequent video, because the audience learns that your openings are marketing.
The fix is structural, not verbal. Write the payoff first, then write the hook that makes it necessary. Hooks written before their payoffs tend to promise whatever sounds best.
A hook that is more interesting than its video is a debt, not an asset.
4. The gap arrives, but late
The hook is good and it is in position four. In front of it sit a greeting, a channel name, a throat-clear and a line establishing that the creator is about to say something. Each is individually defensible. Together they spend the only seconds that were guaranteed an audience.
The fix is mechanical. Find the first sentence that would survive being the only sentence, and delete everything above it. This is the single highest-yield edit in short-form and it costs nothing.
Diagnosing from the data rather than the vibe
These four look different in retention. A video with no gap loses viewers in a smooth early slide. A gap nobody has loses a specific slice immediately — the people who could not place themselves in the topic — and then holds the rest well. A gap that never closes holds viewers deep into the video and loses them at the moment the payoff should have landed, which is the most legible signature of the four. A late gap shows a cliff in the first second or two and a normal curve after it.
If you cannot see retention at that resolution, the proxy is cruder but usable: compare the video's watch-through against your own baseline rather than against other creators, because the absolute numbers are mostly a function of audience size.
What not to conclude
Not every underperforming video has a hook problem. Posts fail because the topic was narrow, because the format fought the topic, or because Instagram reduces distribution for reasons that have nothing to do with the writing — visible watermarks, heavy borders and majority-text frames are all named publicly. Fixing a hook that was never broken is how creators end up rewriting the same opening for a month.